Contactless Asphaltene Detection Using an Active Planar Microwave

Subscriber access provided by McMaster University Library

Article

Contactless Asphaltene Detection Using an Active Planar Microwave Resonator Sensor Mohammad Abdolrazzaghi, Mohammad Hossein Zarifi, Cedric F. A. Floquet, and Mojgan Daneshmand Energy Fuels, Just Accepted Manuscript • DOI: 10.1021/acs.energyfuels.7b00589 • Publication Date (Web): 26 Jun 2017 Downloaded from http://pubs.acs.org on June 28, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Energy & Fuels is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

1

Contactless Asphaltene Detection Using an Active

2

Planar Microwave Resonator Sensor

3

Mohammad Abdolrazzaghi1*, Mohammad Hossein Zarifi1, Cedric F. A. Floquet2, and Mojgan

4

Daneshmand1*

5

1.

6

Canada.

7

2.

8

*: Corresponding Authors

9

KEYWORDS: active microwave resonator sensor, Asphaltene, liquid concentration, solid

Electrical and Computer Engineering Department, University of Alberta, Edmonton

Schlumberger DBR Technology Center, Edmonton, Canada.

10

particles sensing.

11

ABSTRACT

12

In this paper, a noncontact real time method for asphaltenes concentration measurement and

13

precipitation detection is presented. An active high sensitivity microwave resonator sensor is

14

utilized to monitor asphaltene particles and its complex permittivity at 1.2 GHz. The sensor

15

utilizes a microstrip microwave resonator with an active feedback loop to compensate for the

16

environmental loss and to enhance the quality factor of the sensor by several orders of magnitude

17

(>10K). The sensor is utilized to directly measure the concentration of the liquid through

ACS Paragon Plus Environment

1

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 25

1

monitoring its complex permittivity. It is also illustrated that time-based measurements can be

2

used to increase the accuracy by providing two independent data sets: i.e. concentration and time

3

constant. Next, heptane is mixed with different concentrations of asphaltene and monitored in

4

real time. The results demonstrate the capability of noncontact measurement of the precipitation

5

process at different asphaltenes concentrations. The findings exhibit a substantial step toward a

6

low-cost, real-time and noncontact chemical monitoring device.

7 8

1. Introduction

9

Asphaltene precipitation and deposition during production, or transportation of crude oil is a

10

critical issue for the oil and gas industry.1,2 This type of organic scaling can cause plugging of

11

wells or flow lines, reservoir impairment, and more generally, fouling issues necessitating costly

12

remedial maintenance.

13

The asphaltene fraction is the most polar fraction of crude oil and is defined as a solubility

14

class.3 Asphaltenes is soluble in aromatic hydrocarbons, yet, insoluble in alkanes such as heptane

15

or pentane. Pressure, temperature, or chemical composition change in the environment can

16

initiate precipitation. Successful prevention of scaling issues requires the analysis of the

17

asphaltene fraction of a crude oil, as well as a theoretical and practical understanding of the

18

environmental parameters affecting the asphaltene stability. In particular, knowledge of the

19

asphaltene content of the crude oil and pressure onset precipitation can be crucial.1


2

Page 3 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Figure 1. Schematic of microfluidic sensing paradigm 1

Traditionally, asphaltene content is determined with a wet chemistry method.3 Onset

2

(precipitation) determination can be performed using a wide range of techniques: gravimetric,

3

acoustic resonance, or near infrared (NIR) scattering.4 These methods are resource intensive, and

4

operator dependent. VIS-NIR spectrum based techniques study the optical absorption spectrum

5

of the materials and are based on correlations between the sample coloration and the asphaltene

6

concentration in asphaltene/toluene solutions.5–7 Samples ranging between 0.5-15 g/L have been

7

tested with optical spectroscopy techniques.8 While new UV-Vis spectroscopic techniques

8

address some of these issues,9 the nature of the measurement still requires optical access to the

9

sample: optical windows can be prone to fouling and may require tailored engineering solutions

10

and materials to survive exposure to oilfield fluids. A robust and contactless technique, where the

11

optical transparency of the test vessel is not necessary, is sought after.

12

Sensing using microwave resonators is a promising approach to address these challenges (Fig.

13

1). Such sensors rely on measuring changes in their resonant frequency when the environmental

14

dielectric properties vary. Different resonant cavity designs exist with planar systems offering a

15

small size, embedded solution. Planar microwave resonator sensors demonstrated reliable

16

performance in industrial, biomedical and environmental applications.10–14 Recently, passive

17

microstrip resonators, as dielectric property sensors, attracted attention due to their compact size,

18

non-contact sensing capabilities, simplicity, and CMOS (complementary metal oxide


3

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 25

1

semiconductor) and lab-on-a-chip compatibility,15–22 Split ring resonators (SRRs) have been

2

reported as the core of sensing device for material, humidity, temperature, and Nano-structures

3

analyzers.23–27 Lossy dielectric media can also be monitored using SRRs.28,29

4

An important indicator of the performance of a resonator is the quality factor (Q-factor). The

5

Q-factor is defined as the ratio of the resonant frequency to the bandwidth of the resonator. It can

6

also be expressed in terms of the ratio of energy stored over the power loss of the system. The Q-

7

factor plays a critical role in the resolution of the sensor and in the electric field concentration in

8

the sensor medium.22,29–33 High quality factors, a translation of low electromagnetic losses, can

9

increase the resolution of the microstrip resonator sensor and relax the requirement for close

10

physical distance of the medium and the sensor, therefore, enabling contactless measurements.

11

Recently, planar microwave resonator structures with an enhanced quality-factor have been

12

reported for sensing applications, where an active feedback loop compensating for the dielectric

13

loss has been embedded to the conventional structure. The active compensation resulted in Q-

14

factors improving from 200 to over 10k typically.22,29,31–33 The method can potentially

15

compensate the environmental losses and enable high performance sensing in lossy mediums

16

such as liquid. 22,28,29,34

17

In this feasibility study, we investigate the use of an active microwave resonator sensor to

18

detect, measure and monitor dissolved and precipitating asphaltenes. A model oil composed of

19

extracted asphaltenes re-dissolved in toluene is used. We first propose two different schemes to

20

measure the asphaltene content of model oil samples within a wide range of concentrations

21

0.00125-10 % (v:v). First, individual injections of samples inside a fixed tube are tested, and

22

then a diffusivity based method using a reference liquid is implemented to improve the

23

performance. We next demonstrate an active resonator based technique to monitor precipitation


4

Page 5 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

1

and deposition. We present our findings as a proof of principle, a first step towards the

2

development of an instrument capable of replacing optically based measurements. Compared to

3

passive schemes, active sensors, with their loss-compensated resonance and sharp sensing

4

profile, present improved performance: better accuracy for detecting resonance frequencies and,

5

increased resolution allowing the sensing of minute differences between model oils. The

6

proposed sensors with the configurations described in this manuscript are designed to operate at

7

laboratory conditions and could be used for applications such as asphaltene inhibitor screening

8

and selection.

9 10

2. Method and materials 2.1.1. Model oil preparation

11

We used model oils to affranchise ourselves from any potential matrix interference, and to

12

ensure a wide high resolution measurement range. Asphaltenes is extracted from a heavy crude

13

oil following a simplified ASTM D6560 protocol3, where only one rinsing stage is implemented.

14

The stock solution of model oil is prepared by dissolving the extracted asphaltenes in HPLC

15

grade toluene to generate a 25% (w:w) solution. A range of model oils is prepared by volumetric

16

dilution in toluene. To enable direct comparison with existing optical methods, each model oil is

17

diluted again in toluene by a factor 40 and its concentration is expressed in % (v:v) of the initial

18

solution.

19 20

Prepared solutions were kept in a 5 mL borosilicate glass vial (εr=4.3, tan(δ)=0.0047) with a PTFE seal lid.


5

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1 2

Page 6 of 25

HPLC grade heptane (n-C7) was used as a precipitant with a volume ratio of 1:10 (model oil : n-C7) for the precipitation experiments.

3 4

2.1.2. Data acquisition

5

The vector network analyzer (VNA) is used to compute the scattering parameters of the sensor

6

as a device with two ports. Transmission profiles (S21) are collected from a R&S-ZVL VNA

7

(Rohde and Schwarz, Germany) and recorded with LabView (National Instruments, US) at a rate

8

of 1 sample/10sec. The data is then post processed to extract the S21 resonance frequency and

9

amplitude. In cases of mapping the extracted variations in amplitude and frequency over time

10 11 12 13 14

with specific profiles, MATLAB is used for feature extraction. The resonator was kept in a temperature controlled sealed box purged with dried nitrogen gas to prevent the effect of temperature and humidity to interfere with the signal. 3. Results 3.1. Resonator schematic and design


6

Page 7 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

A

Table 1:B Critical dimensions of the sensor Parameter

a

b

I

g

W1

W2

S

Dimension (mm)

26.0

17.0

44.0

2.0

2.1

1.0

0.4

Figure 2. A: Schematic of the active sensor. B: Photograph of the sensing setup (MUT stands for material under test) 1

The concept of the resonator is similar to that of published in previous work.30 Here, the planar

2

resonator sensor is a half wavelength ring resonator magnetically coupled to an input and

3

transmission lines, resulting in a Gaussian peak signal at the frequency of resonance. A

4

schematic of a typical sensor used for this study is shown in Fig. 2A. Its critical dimensions are

5

summarized in Table 1. The strong electric field concentration at the gap ‘g’ of the ring resonator

6

provides the sensitive region for sensing. For a passive resonator, the typical quality factor of

7

less than 300 is limited by losses in transmission. By adding a negative resistor with a feedback

8

loop control to compensate for electromagnetic losses,30 the passive resonator becomes an active

9

sensor with enhanced performance. A NE68033 (CEL, USA) transistor with a 3 GHz gain cut-

10

off frequency is used as an amplifier. To achieve optimum high frequency stability and loss

11

compensation capability, we followed the manufacturer recommendations and operated the

12

transistor at frequencies around 1.2 GHz. Figure 3 shows a model of the sensor simulated using

13

Finite Element Method (FEM) in ANSYS HFSS (ANSYS Inc. US35) with microstrip traces of 35

14

thickness on Rogers 5880 substrates ( 2.2, tan 0.0009). The active circuit is

15

modeled as a negative resistance in parallel with the resonator. The resonance frequency is

16

designed to consider the parasitic effects of the resonator as a shunt capacitance (as small as 50


7

-20

Page 8 of 25

Measurement Full-Wave Simulation G

-40

-60

Sensing Hotspot

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Transmission (dB)

Energy & Fuels

-80 1.1880

1.1913

1.1946

1.1979

Frequency (GHz) Figure 3. Comparison of the measured and simulated response of the planar resonator. The

sensing area is called the hotspot 1

fF) with the ground that is due to imperfections in fabrication and the loading effect of soldering

2

errors. Fine tuning the negative resistance value creates a resonant profile in good agreement

3

with the measured data. For the S21 transmission parameter at a resonance frequency of 1.19135

4

GHz, a good agreement between simulation and measured response is achieved. The quality

5

factor improves from 120 to 10,000 for a passive and active resonator respectively, resulting in

6

an extremely sharp resonant profile

7

3.2. Benefits of an active resonator


8

Page 9 of 25

1

The active resonator used in the sensor provides a very sharp resonance S21 transmission

2

profile such that a high resolution is achievable as shown in Fig. 4A. The figure presents two

3

different profiles which have 1MHz difference. Because of its high quality factor and high roll-

4

off in the skirt of the resonance frequency peak, tracking small changes is possible. On the other

5

hand, the relatively low quality factor of the passive resonator does not allow for a clear

6

distinction in 1 MHz span. The low-quality factor of 120 is prone to amplitude noise, resulting in

7

a high uncertainty in recognizing the S21 resonance frequency. To further demonstrate the

8

improvements brought by the active scheme, the resonant frequencies of the passive and active

9

sensors were measured over 1000 s and plotted in Fig. 4B. A scatter of ± 300 kHz in passive

10

mode compared with less than 10 kHz in active mode were recorded, demonstrating the superior

11

capabilities of the active scheme. 3.3. Concentration measurements

1.121

1.122

MUT-2

1MHz -44

400

-66 -88

-29.00 -29.25

MUT-1

B

MUT-2

1MHz

-29.50

Passive

200 100 0 -100 -200 -300 -400

-29.75 -30.00 1.119

Active

300

∆fres (kHz)

A

-22

1.120 MUT-1

Active

Transmission (dB)

1.119

Passive

12

Transmission (dB)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

0 1.120

1.121

200

1.122

400

600

800

1000

Time (sec)

Frequency (GHz)

Figure 4. A: Illustrative example of the difference of resolution between a passive and an active sensor. MUT: material under test. B: Impact of the active scheme on the precision of the sensor.


9

Energy & Fuels

-10

B

A

Introducing Samples Placing Tube

Transmission (dB)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 25

-20 -30 -40

Sensing Range

-50 -60 -70 1.188

1.191

1.194

1.197

Frequency (GHz)

Figure 5. A: Graphical representation of the concentration measurement scheme. B: Measurement principle. 1

3.3.1. Test setup

2

A schematic of the active sensor and measurement setup is found in Fig. 5A. A 1/8’’ PTFE

3

tube is placed horizontally touching the sensor surface such that it fully covers the hotspot. The

4

position of the tube is kept fixed to the sensor’s surface throughout the entire experiment so that

5

it can be considered as a reference for different materials under test (MUT). The resonator can be

6

modelled as a parallel RLC circuit with Resistance R, Inductance L, and capacitance C, which

7

resonates at fr=1/2π(LC)1/2. Its behavior depends on the effective permittivity of the sensor

8

environment (εr-eff), which can be projected into a capacitance in the circuit model. The bare

9

sensor resonant profile demonstrates a frequency downshift, due to capacitive loading of the tube

10

on the resonator, as illustrated in Fig. 5B. When the material under test (MUT) affects the (εr-eff),

11

the resonant profile varies accordingly, such that higher dielectric constants shift the frequency to

12

lower values, whereas existence of loss in the MUT would tend to shift it upwards. Further

13

details on the theory behind open-ended resonators and the parameters influencing their

14

resonance frequency can be found the Supplementary Information (SI).

15

3.4. Direct concentration measurements


10

Page 11 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

1

A view of the setup is shown in Fig. 5A. A volume of 200-µL of model oil at

2

concentrations of 0.00125, 0.0025, 0.0125, 0.025, 0.1875, 0.3125, 0.625, 1.25, 2.5, and 10 %

3

(v:v) was injected sequentially. In this test, heptane is used as reference material to create a

4

reference data point and eliminate possible environmental and user error from the measurement.

5

Model oils of different asphaltene concentrations are then injected.


11

Energy & Fuels

B

A

From pumps

Sensing Figure 7. A) sensing setup for viscosity measurements and B) schematic of the test for viscosity

1

The measurements are performed in triplicate with the error bar representing the standard

2

deviation of the average is given. As the permittivity of asphaltenes (~5)36 is larger than toluene

3

(~2), an increasing shift in the resonance frequency with increasing asphaltene concentration is

4

observed. The full set of data is plotted in Fig. 6, where the logarithmic scale is used to zoom in

5

the low range of concentrations in inset. The resonance frequency varies from 5.0 MHz to 5.4

6

MHz for the full range of concentrations. A linear fitting function is employed to interpolate the

5.2 ∆f (MHz)

5.08

5

Δf (kHz ) ≅ 53C + 5233

60

5.06

40

5.04 20

5.02

0

5.00

4.8

0.001

0.01

∆f_Offset (kHz)

5.4

∆ f (MHz)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 25

0.1

4.6 0

2

4

6

8

10

12

Concentration (%)

Figure 6. Resonance frequency shifts obtained at different concentrations.


12

Page 13 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

1

intermediate values as shown in Fig. 6. Resolving these frequency shifts is only possible because

2

of the active resonator. As shown before, a passive resonator has a precision of around ±0.3

3

MHz, and would bury the measured signal and information into the noise.

4

3.5. Concentration measurement with a diffusivity based method

5

In this section, the reference is chosen to be “n-Heptane” which fills the tube while acting as

6

a reference. We used the Taylor-Aris dispersion effect to our advantage. Each plug of model oil

7

injected in the tube was followed by a plug of heptane at a flow rate of 0.5 mL min-1. The

8

enhanced diffusivity induced by the shear flow in the tube generated a tail of precipitating

9

asphaltenes. Fig. 7A-B shows the test setup, where a slug a model oil is swapped for a plug of

10

heptane in front of the hotspot of the sensor. Continuous recording of the data is done by

11

LabView37 with a rate of sample/10 sec.

12

Fig. 8A depicts the dispersion of the flow from a complete oil sample as it transits to heptane.

13

Each concentration of model oil has specific dispersion rate, meaning that each concentration

14

will have a different transition time (defined as time constant). Fig. 8B shows the typical

15

frequency shift when a slug of a model oil is swapped for a plug of heptane in front of the

16

hotspot of the sensor. A time series of photographs in Fig. 8B illustrates the process.

17

A


13

Energy & Fuels

Fres (GHz)

1.19140

1.25 % Oil Sample

1.19105 1.19070 1.19035 0

1

20

40

60

80

100

120

140

Time (sec)

B

2

Figure 8. A) Dispersion flow of oil sample in combination with n-Heptane (not to scale,

3

exaggerated to exemplify the measurement principle), B) Frequency shift recorded when pushing

4

a plug of 1.25% (v:v) model oil with HPLC grade heptane.

5

The transition phase between both fluids can be seen at the inflection point of the sigmoidal

6

graph for resonant frequency’s time-based trend, which is called the residence time (τ), and

7

extracted for each model oil samples. Each value is then plotted as shown in Fig. 9A. For

8

comparison, the resonance frequency shifts are also plotted in Fig. 9B. With the diffusivity based

25 20

Residence Time in Dispersion

B

τ = 27 + 4.5 C

1100 840

∆ f (kHz)

A

τ (sec)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 25

15 10

Δf ( kHz ) = 17 C + 842

1000

825 810

900 795

800

5 1E-4 0.001 0.01

0.1

0.5 1 2

5

10

20

780

-2

0

Concentration (%)

2

4

0.001

6

8

0.01

0.1

10 12 14 16 18 20

Concentration (%)

Figure 9. A) Transition time versus model oil concentration. B) Frequency shift versus model oil concentration


14

Page 15 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

Table II, Fitting parameters Parameter

P00

P01

P02

Value

92.12

-3.2

-0.008

Parameter

P10

P11

P20, P12, P21, P22

Value Figure

10.

Relationship

-0.12 0.004

0

between

concentration, resonant frequency shifts and their time constants 1

method, the full range of asphaltene model oil concentration, from 0.00125 to 10% (v:v), can be

2

measured.

3

3.5.1. Dual parameter analysis

4

In order to fully exploit the extracted information in characterizing the diffusion process, a

5

combination of the measured data is presented in Fig. 10, where the logarithmic scale for

6

concentration is assumed to be a function of both measured ∆f and τ. A polynomial function

7

(spatial surface) is pursued to be used as a mathematical formula in predicting intermediate

8

values as follows:

9

∑, Δ

(1)

10

where Pij|i,j=0,1,2 are parameters given in Table II after optimizing the fitting in MATLAB,

11

which yielded a great agreement of R2=0.99 with the experimental data. A 2nd order polynomial


15

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 25

1

function is the simplest case for the analysis to cover all the data, and still include the

2

individual/mutual impact of extracted features of the diffusion process. The two datasets, ∆f and

3

τ, are independent from each other. They convey useful characteristics of the model oils such as

4

the asphaltene concentration and corresponding diffusion rate. Combining the two datasets

5

consolidates the measurement method and reduces the error bars associated with the

6

measurement since it considers a more robust set of information about the sample and not only

7

frequency variations.

8 9 10

3.6. Precipitation detection 3.6.1. Permittivity analysis of suspensions

11

Electromagnetic waves in microwave region interact with materials with respect to their

12

electrical signatures, such as permittivity and conductivity. Mixture of several species (solutions,

13

suspensions, electrolyte, etc.) represents mixed properties of all of its individual constituents

14

according to their volume fractions. As a first order approximation, asphaltenes can be

15

considered as colloids suspended in crude oil and solvable in toluene. Based on DeLacey and

16

White theory,38 the real and imaginary parts of the dielectric constant of a suspension can be

17

calculated if the induced dipole of the suspended inclusion is known:

18



ε '(ω) = ε 'S + 3ϕε 'S C1 (ω) − 

19

 C2 (ω ) ωε 0ε 'S  K∞

 K∞

 (C1 (ω ) − C1 (0)) + C2 (ω )   ωε 0ε ' S 

ε ''(ω ) = 3ϕε ' S 

(2)

(3)


16

Page 17 of 25

B

A

Gain h D2 Hot-spots

C 0

Transmission (dB)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

-10

P

-20

it cip re

on ati

-30 -40 -50

D1

-60 1.45

1.50

1.55

1.60

Frequency (GHz)

Figure 11. A) Schematics of the modelled sensor with dimensions of D2 = 6 mm, h = 20 mm, and variable D1, B) Details of the precipitation model used for the simulation. C) Simulated signal output from HFSS 1

where is the volume fraction of solids, εr' and are dielectric constant and DC conductivity

2

of the host medium, and C1 and C2 are the real and imaginary parts of induced dipole

3

coefficient.39 Here, not only the permittivity of the inclusions and the host medium should be

4

considered separately, but also the volume of liquid that is being substituted by solid particulates

5

is of great importance. Reducing the external inclusion in the mixture with high εr' decreases the

6

effective and inhomogeneous permittivity of the mixture. Therefore, it is expected that by

7

monitoring the permittivity variation in the liquid, one can characterize the mixture, and its

8

precipitation status with regards to time and quantity.

9 10

3.6.2.

Sensor design and simulation

11

In this occurrence, the sensor is a meandered shape half-wavelength resonator on microstrip

12

that is coupled to an active gain-controlling section to enhance its quality factor from ~270 up to

13

~10 k. The operating concept is very similar to the one explained in section 3.1. Details are given

14

in the reference17 and a schematic can be found in Fig. 11A.


17

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 25

1

The resonator has two main hotspots at both of its end parts where maximum electric fields

2

are confined. For the ANSYS HFSS simulation, a cylindrical sample with dimensions of D2 = 6

3

mm, h = 20 mm is placed 1 mm off the left hotspot with εr=2, tan(δ)=0.01 (representing the host

4

medium). To simulate the precipitation, an external material is put inside the host cylinder ( εr=5,

5

tan(δ)=0.1) of variable diameter D1 to represent precipitating materials.

6

asphaltene model gets larger in diameter and shorter in height to mimic the deposition process

7

(Fig. 11B). Results computed using the finite-element method in ANSYS HFSS are shown in

8

Fig. 11C. The top section of the sample is the main part for the sensor. Change in effective

9

permittivity of the environment is represented by deposition where the extracted material (high

10

εr) is moving away from the hotspot and settling down at the bottom of the cylinder. Thus, the

11

high permittivity of the solute is affecting less and less the resonant frequency (fr) over time

12

resulting in an increasing fr.

13

3.6.3.

The precipitated

Precipitation detection


18

Page 19 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

1

The experimental setup is shown in Fig. 12A. The active sensor is mounted vertically in order to

2

monitor the precipitation inside a 5 mL borosilicate glass vial containing the mixture of model

3

oil and heptane. The vial is mounted 5 mm away from the sensor, and the glass thickness is

4

approx1 mm. Three samples with concentrations of 0.1875% (v:v), 0.6250% (v:v) and 1.25%

5

(v:v) and a toluene blank were all mixed with heptane as the precipitant element. A 6-minute

6

time lapse of the deposition of the precipitated asphaltenes for the 0.625% sample is shown in

7

Fig. 12B. Conceptual variation of the size and volume of the aggregated molecules is presented

8

in Fig. 12C, where first nanoparticles start being precipitated. Next, their collection brings about

9

nanoaggregates, in smaller scale, and finally this trend continues in larger scale to result in

A

B

C Figure 12. A) Experimental setup for time-based monitoring of deposition. B) Evolution of the precipitation and deposition over time for the 0.625%-sample (ratio 1:10). C) Conceptually demonstrating precipitation and aggregation


19

0

2

Time 0 (min)

6

4

-20

-40

-60 1.54294

A

1.54298

1.54301

Shift in Resonant Frequency (kHz)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Transmission (dB)

Energy & Fuels

150

1.25 %

100

0.625 % 50

0.1875 % Toluene

0 0

1.54305

Frequency (GHz)

Page 20 of 25

1

2

B

3

4

5

6

7

8

9

Time (min)

Figure 13. A) Transmission profile from the time based deposition in experiment. B) Variation in resonant frequency for four samples. The solid and dotted lines represent two different measurements. Data is normalized to the starting time. 1

clusters that tend to deposit from their weight.

2

Following the same time line, S21 transmission responses of the sensor are shown in Fig. 13A. A

3

Shifts in resonant frequencies are extracted after post processing the results and plotted in Fig.

4

13B. Duplicate measurements were performed. As the particulates of asphaltenes with higher

5

permittivity deposits to the bottom of the glass container, leaving only liquids with lower

6

permittivity in the sensor’s sight, the effective dielectric constant is reduced, which results in

7

shifting fr upwards. The four different sample concentrations present quite distinctive profiles

8

indicating the feasibility of quantifying the deposition after calibration.

9

4. Conclusion

10

We demonstrated that a high Q-factor high-resolution microwave resonator based sensor can

11

be used to successfully detect and quantify the asphaltene content of a set of model oils between

12

0.00125% (v:v) and 10% in noncontact fashion. A calibration curve is required to quantify the


20

Page 21 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

1

asphaltene content of the sample. Time based measurement is also used to obtain independent

2

information and measure time constant and permittivity in determining low concentrations of oil.

3

It is also shown that such a system can detect and monitor the asphaltene precipitation and

4

deposition of a model oil when mixed with heptane in real time. The sensor has a small form

5

factor, is compatible with microelectronic manufacturing processes, and can be integrated into

6

small device. Further work will focus on porting the method to crude oil samples.

7

Corresponding Authors

8

* Mohammad Abdolrazzaghi: Phone: +1-587-938-3690; email: [email protected]

9

* Mojgan Daneshmand: Phone: + 1-780-492-7351; e-mail: [email protected]

10

Funding Sources

11

The research is funded through National Sciences and Engineering Research Council (NSERC)

12

and Alberta Innovates Futures (AITF).

13

ACKNOWLEDGMENT

14

The authors would like to acknowledge Canada Research Chair, AITF, CMC microsystems, and

15

NSERC for providing funding and Rogers Corporation for generous substrate donations. We

16

thank Schlumberger Canada Ltd for their in-kind contributions.

17 18 19

5. References (1)

20 21

Frenier, W. N.; Ziauddin, M.; Venkatesan, R. Organic Deposits in Oil and Gas Production.

(2)

Sheu, E. Y. Energy & Fuels 2002, 16 (1), 74–82.


21

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

(3)

Page 22 of 25

PA, 2012 ASTM D6560-12 Standard Test Method for Determination of Asphaltenes

2

(Heptane Insolubles) in Crude Petroleum and Petroleum Products. ASTM Int. West

3

Conshohocken, PA,.

4

(4)

Jamaluddin, A. K. M.; Creek, J.; Kabir, C. S.; McFadden, J. D.; D'Cruz, D.; Joseph,

5

M. T.; Joshi, N.; Ross, B. In SPE Asia Pacific Improved Oil Recovery Conference; Society

6

of Petroleum Engineers, 2001.

7

(5)

Ruiz-Morales, Y.; Wu, X.; Mullins, O. C. Energy & fuels 2007, 21 (2), 944–952.

8

(6)

Mullins, O. C.; Betancourt, S. S.; Cribbs, M. E.; Dubost, F. X.; Creek, J. L.; Andrews, A.

9 10

B.; Venkataramanan, L. Energy & Fuels 2007, 21 (5), 2785–2794. (7)

11 12

2001, 15 (4), 979–986. (8)

13 14

Ching, M.-J. T. M.; Pomerantz, A. E.; Andrews, A. B.; Dryden, P.; Schroeder, R.; Mullins, O. C.; Harrison, C. Energy & Fuels 2010, 24 (9), 5028–5037.

(9)

15 16

Joshi, N. B.; Mullins, O. C.; Jamaluddin, A.; Creek, J.; McFadden, J. Energy & Fuels

Schneider, M. H.; Sieben, V. J.; Kharrat, A. M.; Mostowfi, F. Anal. Chem. 2013, 85 (10), 5153–5160.

(10)

17

Schueler, M.; Mandel, C.; Puentes, M.; Jakoby, R. IEEE Microw. Mag. 2012, 13 (2), 57– 68.

18

(11)

Gregory, A. P.; Clarke, R. N. IEEE Trans. Dielectr. Electr. Insul. 2006, 13 (4), 727–743.

19

(12)

Zarifi, M. H.; Shariaty, P.; Abdolrazzaghi, M.; Hashisho, Z.; Daneshmand, M. Sensors

20

Actuators B Chem. 2016, 234, 332–337.


22

Page 23 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

Energy & Fuels

(13)

2 3

Chem. 2017. (14)

4 5

(15)

(16)

(17)

Zarifi, M. H.; Fayaz, M.; Goldthorp, J.; Abdolrazzaghi, M.; Hashisho, Z.; Daneshmand, M. Appl. Phys. Lett. 2015, 106 (6), 62903.

(18)

12 13

Withayachumnankul, W.; Jaruwongrungsee, K.; Tuantranont, A.; Fumeaux, C.; Abbott, D. Sensors Actuators A Phys. 2013, 189, 233–237.

10 11

Abdolrazzaghi, M.; Abdolali, A.; Hashemy, S. Appl. Comput. Electromagn. Soc. J. 2016, 31 (2), 126–131.

8 9

Rydosz, A.; Maciak, E.; Wincza, K.; Gruszczynski, S. Sensors Actuators B Chem. 2016, 237, 876–886.

6 7

Staszek, K.; Rydosz, A.; Maciak, E.; Wincza, K.; Gruszczynski, S. Sensors Actuators B

Abduljabar, A. A.; Rowe, D. J.; Porch, A.; Barrow, D. A. IEEE Trans. Microw. Theory Tech. 2014, 62 (3), 679–688.

(19)

14

Abduljabar, A.; Xin Yang, X.; Barrow, D.; Porch, A. In 2015 IEEE MTT-S International Microwave Symposium; IEEE, 2015; pp 1–4.

15

(20)

Naqui, J.; Martin, F. IEEE Trans. Microw. Theory Tech. 2013, 61 (12), 4700–4713.

16

(21)

Rusni, I.; Ismail, A.; Alhawari, A.; Hamidon, M.; Yusof, N. Sensors 2014, 14 (7), 13134–

17 18 19

13148. (22)

Abdolrazzaghi, M.; Zarifi, M. H.; Daneshmand, M. IEEE Sens. J. 2016, 16 (22), 8151– 8157.


23

Energy & Fuels

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 25

1

(23)

Lee, H.-J.; Lee, H.-S.; Yoo, K.-H.; Yook, J.-G. J. Appl. Phys. 2010, 108 (1), 14908.

2

(24)

Horestani, A. K.; Fumeaux, C.; Al-Sarawi, S. F.; Abbott, D. IEEE Sens. J. 2013, 13 (4),

3 4

1153–1160. (25)

5

Galindo-Romera, G.; Javier Herraiz-Martinez, F.; Gil, M.; Martinez-Martinez, J. J.; Segovia-Vargas, D. IEEE Sens. J. 2016, 16 (10), 3587–3596.

6

(26)

Boybay, M. S.; Ramahi, O. M. IEEE Trans. Instrum. Meas. 2012, 61 (11), 3039–3046.

7

(27)

Lee, C.-S.; Yang, C.-L. IEEE Sens. J. 2014, 14 (3), 695–700.

8

(28)

Zarifi, M. H.; Daneshmand, M. Sensors Actuators B Chem. 2016, 225, 517–521.

9

(29)

Abdolrazzaghi, M.; Daneshmand, M. In 2016 IEEE MTT-S International Microwave

10 11

Symposium (IMS); IEEE, 2016; pp 1–3. (30)

12 13

230. (31)

14 15

(32)

Abdolrazzaghi, M.; Zarifi, M. H.; Daneshmand, M.; Floquet, C. F. A. In SENSORS, 2016 IEEE; IEEE, 2016; pp 1–3.

(33)

18 19

Abdolrazzaghi, M.; Zarifi, M. H.; Pedrycz, W.; Daneshmand, M. IEEE Sens. J. 2017, 17 (2), 323–332.

16 17

Zarifi, M. H.; Thundat, T.; Daneshmand, M. Sensors Actuators A Phys. 2015, 233, 224–

Abdolrazzaghi, M.; Zarifi, M. H.; Daneshmand, M. In SENSORS, 2016 IEEE; IEEE, 2016; pp 1–3.

(34)

Zarifi, M. H.; Daneshmand, M. In 2015 IEEE MTT-S International Microwave


24

Page 25 of 25

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Energy & Fuels

1

Symposium; IEEE, 2015; pp 1–4.

2

(35)

ANSOFT HFSS® Academic Research, R. 2016. .

3

(36)

Evdokimov, I. N.; Losev, A. P. Energy and Fuels 2010, 24 (7), 3959.

4

(37)

National Instruments Website. http://www.ni.com/labview/.

5

(38)

Jiménez, M. L.; Arroyo, F. J.; van Turnhout, J.; Delgado, A. V. J. Colloid Interface Sci.

6 7

2002, 249 (2), 327–335. (39)

Goual, L.; Firoozabadi, A. AIChE J. 2002, 48 (11), 2646–2663.

8


25

Contactless Asphaltene Detection Using an Active Planar Microwave

Recommend Documents